When you start a DEIMOS science exposure, it's critical that the
Flexure Compensation System (FCS) be in Tracking mode
when the exposure starts. If the exposure begins before FCS has
settled then the image will move on the detector during the
first few minutes of the exposure, resulting in image blur.
Instead of simply clicking the Expose button on the
DEIMOS dashboard to begin an exposure, we suggest that you enter
the following command into a polo xterm window:
wffcs ; goi
This will cause the system to wait until FCS is tracking, then
begin the exposure. You can run this command even while
reconfiguring DEIMOS, since the FCS will not be tracking until
the instrument is done being reconfigured. Thus, you can run
this command as soon as your slitmask alignment is complete and
you are sending the grating back from the zeroth order imaging
position to the science wavelength.
Although the guider software does not measure a true FWHM
parameter, we have a handy utility for measuring the seeing off
of slitmask alignment images. From within IRAF, simply type the
command get_seeing; this script will fit Gaussian
profiles to the stars in your alignment boxes and compute and
print a median seeing value.
We recommend that you verify the alignment of your slitmask by
running the align_check
script regularly during your exposure sequence. Open a
polo xterm window and initiate the
align_check command while a spectral exposure is in
progress. The script will wait for the exposure to complete,
then reconfigure the grating to the imaging position. It will
then acquire a single alignment image and return the grating to
the science wavelength so that FCS can re-establish lock. All
you need to do is run do_xbox in IRAF (do this after
the spectral image has read out) and analyze the resulting
direct image. Send any required moves and then press the
<Enter> in the xterm running
align_check to initiate the next exposure.
One possible reason is that you changed gratings; the mapping
from position on the slitmask to position on the CCD image
depends on the grating; thus, when you change gratings you need
to re-run the tune_qmodel program to allow
xbox to locate your alignment boxes.
However, incorrect observing strategy can also require you to
recalibrate xbox. Although it is tempting to
increase your observational efficiency by reconfiguring DEIMOS
during a telescope slew, this strategy can be
counterproductive. Since the DEIMOS FCS cannot track when the
instrument is being reconfigured, you prevent the FCS from
doing its job right when the most flexure is occurring! The
typical result is that when you get onto your new target and
acquire a direct image for slitmask alignment, you find that
the xbox software is unable to locate your
alignment boxes due to the large uncorrected flexure that
occured during the slew. If you then run the
tune_qmodel task to update the mapping from mask to
CCD coordinates, you will fix the problem...but only
temporarily. The next time you go to align a mask, the FCS
will now have updated the flexure corrections and hence your
updated coordinate mapping will be incorrect again, forcing
you to run tune_qmodel yet again!
The preferred method for dealing with this is to allow DEIMOS
to remain in its spectroscopic configuration during the
telescope slew. You can complete the coarse alignment phase
of the slitmask alignment (through running gmov) while the FCS is making
corrections. Once the FCS status changes to
Tracking, click the GO ZERO button to
switch to imaging mode and acquire your exposures for running
xbox. This should ensure that:
Yes. The tune_qmodel task can only store data for
one grating, and since the different gratings require different
parameters for the optical model you will generally need to
re-run the tune_qmodel task. For maximum efficiency,
we recommend that you re-run tune_qmodel
before you start the mask alignment process on the new
grating; simply run tune_qmodel on any direct imaging
exposure which already exists on disk using the new grating.
For example, if you acquired direct images of any mask with the
new grating during the afternoon calibrations, these can be used
as input to tune_qmodel. Doing this should ensure
that your optical model is correctly set and thus that
xbox can locate your slitmask alignment boxes
Most DEIMOS stages can be moved during CCD readout without
compromising the CCD readout noise. However, the dewar focus
mechanism and the dewar filter wheel mechanism will inject a few
electrons of fixed pattern noise if they are moved while the CCD
is reading out. This is due to the fact that the motors for
those mechanisms are in extremely close proximity to the DEIMOS
CCD electronics boxes and associated cabling.
If xbox fails to find certain boxes on a mask,
inspect the alignment image on ds9 to see whether
some regions of the image are badly out of focus. If so, try
the following steps to fix the mask:
- xbox can find your alignment boxes without
needing to re-run tune_qmodel, and
- the FCS will rapidly return to Tracking mode
once you complete the slitmask alignment process and return
to spectroscopic mode.
NOTE: see also the Slitmask
Database FAQ page.
The minimum slit width is 0.63 arcsec for a slit which is untilted
(i.e., the position angle of the slit matches the position angle
of the mask).
The answer depends on your slit width, according to the formula
- Re-insert the mask by clicking the Retract
Mask button on the Slitmask Panel and re-selecting
the mask, then take another image and check the slits.
- If this fails, have the OA rotate the instrument to a
drive angle of +60, then retract and re-insert the mask and
take an image to check the slits. Try as many times as
your patience allows.
- If this fails, switch to another mask.
pa_max = acos(0.63/w)*pi/180.
where pa_max is the maximum PA of the slit (relative to
the mask), and w is the slit width in arcsec. For a
slit width of 1.0 arcsec, the slit may be tilted by up to
50° relative to the mask position angle. For a slit width of
0.7 arcsec, the maximum tilt is 26°.
Yes, we have written an IRAF utility called plotmask which converts these files into
a format compatible with the STSDAS IGI graphics program and
plots them. Please feel free to download and use the utility.
Elementary, my dear slitmask observer! Proceed as follows:
Please note that multiple objects could be assigned to each slit.
We have seen this happen in several instances in which mask
designers used the t (toggle) function in DSIMULATOR to
change program objects (defined as targets with priority codes
greater than zero) into guide stars. The t key should
only by used to switch stars (i.e., objects with priority
codes of -1 or -2) between alignment and guide star.
- Note the BSLIT value indicated in the error
- Log in to the slitmask database
- Click on the button Show Me My Mask Inventory
- In the mask list, locate the mask of interest and in the
rightmost column click the Detail button to get
the details page.
- In the BluSlits table, search the
bSlitId column for the BSLIT number
you are seeking.
- In the dSlitId column, note the corresponding
- In the SlitObjMap table, search the
dSlitId column for the DSLIT value.
- Note the OBJECT value in the
- Click the hyperlink under ObjectId to go to
the corresponding entry in the Objects table.
- The entry in the Objects table indicates the
name and characteristics of the object in the slit.
To fix this, you can simply re-generate your slitmasks by using
the text files generated by DSIMULATOR as input. The text file
will correctly list the priority codes for the guide stars as
-2, and so when you use this as input to DSIMULATOR it will
generate appropriate alignment boxes for these targets.
The answer depends on which grating is to be used. The DEEP2
pipeline seems to work better with certain grating/slider
combinations than others:
Yes. Due to the large acceptance angle of the DEIMOS camera,
all gratings may produce zeroth-order ghosts at some
grating tilts. The 830G and 900ZD gratings are particularly
prone to this at typical grating angles. These gratings are
certainly usable with slitmasks, but observers using these
gratings should be prepared for ghosts in their data. Please
read the notes on ghosting with the
900ZD grating and heed the advice on how to acquire
calibrations with the 830G and 900ZD gratings.
No. Variations of order 10% are observed from the top to the
bottom of the detector, probably as a result of changes in
anamorphic factor with slit position.
Unfortunately, no. The LRIS filters (241×241mm) are much larger
than the DEIMOS filters (165×165 mm) and so the former would not
fit in the DEIMOS holders. Also, the beam in DEIMOS is quite
fast (f/1.29) and thus the existing LRIS narrowband interference
filters would probably give signficant passband variations
across the FOV.
DEIMOS filters must measure (165×165 mm) in size to fit
into the filter holders and be 6.0±0.1 mm thick in order
to allow the camera to focus. Please refer to DEIMOS drawing
5303 for dimensions and tolerances.
The original DEIMOS V-band and BAL12 filters were damaged in 2004 and
have significant scratches. We expect that this will result in
a transmission function that changes across the DEIMOS field of
view, and for that reason we do not recommend using these
filters. Please contact your assigned Support Astronomer if you
still wish to use either of these filters on your DEIMOS run.
Images from the DEIMOS mosaic camera are stored in a format
FITS (as opposed to the “simple FITS” format
employed on earlier-generation Keck instruments such as the
original HIRES, LRIS, ESI, etc.) and must be opened in a
different manner in order to read the entire image into ds9. Instead of
simply using using the option
- 600ZD is preferred in slider 3.
- 830G is preferred in slider 4.
- 900ZD works well in either slider 3 or slider 4.
- 1200G is preferred in slider 3.
File > Open
which will display only the first image extension, you'll need to use
File > Open Other > Open Mosaic IRAF
to view the entire image.
For IDL, please see the program deimos_read_file.pro
distributed as part of the DEEP2
DEIMOS data reduction package. In IRAF, see the MSCRED
package. The Keck IRAF
package includes a task called deimos.dmosdisplay
which uses mscred.mscdisplay to display the full
DEIMOS image on the IRAF image display client.
Likely the deiccd keyword "autopane" was set to an incompatible
value (FullFrame, MaskArea, AlignArea) and not Manual.
For complete info on autopane, please see the Autopane Webpage.
- Set autopane to manual via the dashboard or modify -s
deiccd autopane=manual from a deimosserver xterm.
- Re-set your binning and window mode
- Take your images